Also known as Foreground Detection is a widely known technique to extract image's foreground
This project implements the 1-D Gaussian background subtraction algorithm. Each individual pixel is modeled using a Gaussian distribution.
- Binary thresholding is applied by setting foreground to 255 (white) and background to 0 (black).
- Initial threshold (k) and temporal window (𝜌) values are set based on trial and error.
Instructions to Run
- Install and configure OpenCV.
- Download image sequence from here: Office Dataset
- Extract images into Images directory. (Image file names format: inxxxxxx.jpg)
- First 619 images are used for generating binary threshold version of image in000620.jpg. Adjust variables as need be.
- Compile and run.